FAQ No. 217: The ideal way to protect against surge events in Ethernet and industrial applications
INDUSTRIAL LCD DISPLAYS / IGBT MODULES DISTRIBUTOR

Infineon / Mitsubishi / Fuji / Semikron / Eupec / IXYS

FAQ No. 217: The ideal way to protect against surge events in Ethernet and industrial applications

Posted Date: 2024-01-16

作者:James Niemann, Field Applications Engineer
James Niemann, Field Applications Engineer

question:

Is there an easy way to protect an Ethernet network from lightning damage?

Answer:

With a strong understanding of magnetism and circuit theory, coupled with proper grounding and shielding techniques, a solution can be found.

Summary

Taking appropriate precautions can prevent lightning strikes from damaging Ethernet connected equipment. The traditional approach of using protective components may not be entirely effective, and we need to supplement it with another approach, inspired by an in-depth analysis of the underlying mechanisms of lightning strike energy transfer to Ethernet cables and connected devices, which will be described in detail in this article.

introduction

Surge events with lightning strikes as the culprit can cause wired Ethernet network failures, which always affects the nerves of network administrators or other relevant persons in charge. This problem is not limited to Ethernet, but to any larger real-world electronic or power system. Other examples include: electrical measurements returned remotely, power transmission, and industrial automation applications where the sensors are not nearby. Traditional solutions work by absorbing or limiting event energy near an area to protect physical layer components. The problem with this approach is that the energy is not eliminated and neither is the resulting current. Transient currents in the sensing path always produce large voltages that can cause damage. Therefore, when considering conventional approaches, we must be clear: What level of protection is required, and how much time, effort and resources will be required to implement it? Additionally, the protection methods deployed must not only be able to withstand the surge, but also function properly after the surge occurs. The energy generated by a lightning strike is orders of magnitude greater than you might think. In order to achieve safe and reliable operation, strong protective measures need to be taken to meet this challenge.

Energy size

Ethernet systems need to protect against surge events, and the energy of different surge events can vary widely. An energy surge caused by a lightning strike thousands of meters away may be five orders of magnitude less powerful than an energy surge caused by a lightning strike outside your door. The size of an Ethernet system also affects its ability to handle different amounts of energy. Even the orientation of the loop may increase the surge capacity of the system by three orders of magnitude.

Lightning strike energy

The damage caused by a surge event depends on the energy of the event, where the energy surge occurs, and how much energy the system can store when struck. Understanding these factors will help find solutions to prevent this damage.

The energy generated by a lightning strike is stored in the area surrounding the strike (our discussion excludes the possibility of a direct lightning strike). The main problem with lightning strikes is the storage of energy in the near field, where the magnetic field is most important for this low impedance source. The total inductance can be found from the length of the lightning strike, then using the familiar energy equation E = 1/2Li2The total energy in the magnetic field can then be approximately calculated. Lightning currents vary in magnitude but may be as high as 50,000 A. In the far field beyond this distance, the energy is very small and of little concern unless you are in the business of building radio receivers.

The sun produces 3.846 × 1026 W of power per second. On Earth, 93 million miles from the sun, one square meter of space receives 1,000 W of that power. If we integrate over the entire sphere around the Sun, no matter how far away it is from the Sun's surface, there is always 3.846 × 1026 W of radiated power, and 1 square meter is very small relative to the total surface area 93 million miles away! Now let's talk in terms of energy rather than power. To obtain 1000 joules of energy, 1 s of irradiation is required (watts are measured in J/s). This energy volume is equal to 1M2 times the distance that light travels in one second, which is 3 × 108 M; in this case, the total volume is also 3 × 108M3

In order to understand the rest of this article, one must accept the concept that both radiant energy and static energy (magnetic energy BxH and electrostatic energy ExD) are stored in space. Poynting's theorem describes the motion, transfer, or transfer of energy. The transfer of energy always involves both magnetic and electric fields. There cannot be a significant electric field inside a conductor, so it is impossible to store any energy. It's simple and clear that both near and distant (radiated) energy is stored in the space surrounding the lightning strike event. This concept (energy stored in space) suggests the following solution to the surge problem: Without exposure to this energy, the surge problem is eliminated.

To access this energy, the conductor geometry (Ethernet cable) needs to enter the space where the energy is moving. Just like our radiation example, even in the near field, time is involved. Ethernet cables are connected differentially and do not have any significant loop area, so it is unlikely to couple any significant energy from this surrounding space. This is not the case for the area between the Ethernet cable and the grounding system.

A surge is a high-frequency loop current involving the chassis grounding system. Every circuit built has a chassis grounding system. For the purposes of this article, it is only important for large circuits. See the example in Figure 1 to see how a chassis grounding system is always present, and the larger the system, the more important it becomes, and why grounding has nothing to do with this problem, while any parasitic conductors do. The next section describes the two most common sources of inrush current.

Ground loop energy

Ground loops occur because the ground potential is not constant at any two locations. Figure 1 shows that every schematic has a second circuit, a parasitic ground loop. Because a ground loop and the circuit you design can share a wire, the ground loop is also called common impedance coupling1. Figures 1 and 2 show more detailed examples. Typically, the second chassis ground circuit is not that large, but is always present. Generally speaking, the further the distance an electronic system covers, the greater the potential difference between these grounds, and the greater the inductance and resistance between them.


Figure 1. Technically speaking, even a system as small as a handheld device can be affected by outside influences. In this example, the ground loop is very small and any interference current will flow to the shield instead of the radio ground.


Figure 2. The above figure shows a line-powered instrument with a ground loop voltage between chassis ground 1 and chassis ground 2. At the same time, the loop is large enough for magnetically coupled interference to become significant. Also note that the interference loop shares a wire with the instrument ground.

When lightning strikes the ground, the current spreads in all directions. This current causes a significant voltage drop across the ground resistance and inductance through which the current flows. For some wired Ethernet installations, this potential difference can span the entire Ethernet cable (from one end to the other) and can cause large currents to flow. This effect is classified as a ground loop, which is correct. Currents originating from instrumentation and electrical machinery can also cause ground loops. A properly grounded building is based on a single grounded conductor at the utility entrance. It can be seen that within a single building, ground loops caused by lightning strikes are not the dominant factor causing equipment damage. This is obviously not the case for Ethernet networks run outside or between buildings.

Regardless of the source, ground voltage will produce current in an Ethernet cable, even if the length is not very long or there is no loop area. What matters is the potential difference between the two ground points, the rise time, and the inductance of the chassis system between the two points.

magnetic field produced by lightning

According to Faraday's law, lightning can also produce magnetically coupled voltages in any loop region. This is probably the most worrisome issue because it affects the Ethernet network deployed within the building.

Lightning surge events caused by ground loops are different from surge events caused by magnetic coupling (Faraday's law). The next few sections discuss possible solutions to each problem. For reference, Figure 3 shows an Ethernet connection without any of the solutions described in this article. Here, the current path (due to ground loops or Faraday's law) through the circuit and the ground reference plane (which is also part of the circuit) is the only path through which the surge current can travel. Conventional solutions attempt to shunt this current away from the component, but dangerous V = Ldi/dt events can occur in the current path.


Figure 3. An example of an Ethernet network susceptible to surge damage.

textbook solution

The textbook solution to ground loop and magnetic energy problems is to use guards. Provide protection by surrounding the entire application with a shield. The purpose of this guard is to minimize the capacitance between the application and anything other than the guard itself. Considering Figure 4 (a simplified non-Ethernet example), it is obvious that any ground loop or magnetically induced current will flow along the guard metal and across the isolation barrier at C5. Ground loop currents are unlikely to enter any application area enclosed by the protective device. In this case, the interference field is completely external to the application component. In addition to eliminating any electrostatically coupled noise, this textbook solution also addresses both sources of interference mentioned above. Textbook solution is really great, it works even though the C5 is very small. Short circuit turns are not necessary.


Figure 4. Instrument example showing the use of guards to remove energy and thereby reduce inrush current in an application circuit.

This is the only solution that works for both ground loop and magnetic coupling field energies. It is also generally beyond the requirements of this type of Ethernet application, and some simplifications will be made below to get to an Ethernet solution that we can actually build.

Short circuit turns

The energy that causes damage comes from the energy field created by lightning. In order to eliminate the energy in the Ethernet cabling we need to eliminate the energy field and for this we will design a short circuit turn in this transformer where the lightning is the primary and the Ethernet ground loop area is the secondary. If we use the shield inside the Ethernet cable and the plane in the application circuit to build an isolated shorted turn, with a ground providing the final conductor to close the shorted turn, then we should be able to eliminate the energy. In practice, the process of adding external shunt protection elements is much easier after implementing such short-circuit turns.

To simplify, the complete surround of the left and right halves of the system can be removed, as shown in Figure 5 (the Ethernet configuration is shown in Figure 6). This simpler configuration may be effective if the guard loop can act as a short-circuit turn and the C3/C4 ratio is extremely small. Compared to an isolated path, this simplified method of eliminating surges only works if we are able to construct short-circuit turns.


Figure 5. Simplified instrument example using shielding to divert surge energy away from the application circuit.


Figure 6. Ethernet example using shielding to divert surge energy away from the application circuit, C3

From an Ethernet loop perspective, how exactly does this short turn dissipate energy? To uncover this problem, we need to understand the transformer analogy at a deeper level. Real transformers are designed to move energy, not store it. This is true whether it is an air core transformer or a magnetic core transformer. In order to store almost no energy in an air-core transformer, the windings must be wound directly on top of each other so that there is little room to store energy. Even if the windings are not directly on top of each other, a transformer made with a magnetic core will transfer energy (with hysteresis and eddy current losses) from one winding to the other, but there must be almost no space between the windings and the core so that there is almost No energy is stored. When using magnetic cores, a larger µr directly reduces the magnetizing current, giving us an additional advantage due to the higher inductance. With or without a core, a voltage applied to the primary produces a current, described by the familiar relationship V = Ldi/dt, which in turn causes a voltage on the secondary, given by:
V = (loop area) dB/dt. The presence of magnetic material does not change primary Ldi/dt or secondary dB/dt. In other words, it does not change the transformer voltage. In the primary, the permeability µr is a constant which increases the inductance (increases µr) but also reduces di/dt to compensate. For the secondary, a larger µr will slow down dB/dt (because the primary di/dt is lower), but it will also increase B by that constant. High permeability really just reduces the magnetizing current by increasing the primary inductance.

Since there is no energy stored in the transformer, when the secondary load is large, the primary driven by a low impedance voltage source will need to provide more current, and the primary current will increase to provide energy.

In contrast, a lightning strike stores a large amount of energy in a very large space. Energy always arranges itself in configurations that store as little energy as possible. This is exactly what the transformer does internally and at the secondary winding interface, with the secondary current flowing in the opposite direction to the primary current. These opposing currents ensure that there is no net external magnetic field (stored energy). At a high level, this is called the principle of least action, but for the purposes of this article, it's called Lenz's law. This is what happens in the space around Ethernet cables and chassis ground loops. Ethernet loops (or short-circuit turns, your choice) provide a means to divert or dissipate this energy, since either means provides a way to store less energy. Just like the transformer example above, the resulting secondary voltage is still V = (loop area) dB/dt, but there is no tight coupling between the primary (lightning) and secondary (Ethernet loop). This poor coupling leaves the area inaccessible to unlimited energy sources. Shorting a wire turn creates a current that cancels/dissipates the energy stored in that space by lightning. If the inductance of the primary could be measured with the shorted turns in place, it would be a lower value, indicating less energy was stored and some of the lost energy was dissipated in the shorted turns. In other words, the magnetic field produced by the secondary load will cancel out the magnetic field produced by the lightning, causing less energy to be stored in the Ethernet loop.

By the way, in a transformer, when one secondary is short-circuited, this is exactly what happens. However, there is an important difference. With a real transformer, the shorted turns will dissipate all available energy in the primary due to tight coupling. With lightning, only the energy in the Ethernet ring space is dissipated.

Let's look at an example. The H field generated by a lightning strike is I/2πR. Assuming the lightning strike is 1 mile (1600 M) away from the Ethernet cable and the lightning strike current is 50,000 A, the magnetic field strength will be 4.97 A/M

The B field is B = µH = (4π × 10E-7)(4.97) = 6.25E-6 Tesla. The Ethernet loop area (one mile away) is: 1 M × 150 M =150M2

The rise time of lightning current can be as short as 1μs, and its fall time is about 100 μs, so the voltage generated in this loop can be approximately calculated as: V = A (loop area × dB/dt) = 150(6.25E-6 )/1 μs = 937 V

We perform simulations to obtain accurate values. Figure 7 shows a 50 kA lightning strike with a rise time of 1 μs and a fall time of 10 μs.


Figure 7. 50 kA lightning strike with rise time 1 μs and fall time 10 μs.

According to Faraday's law, this current will produce voltage V1, as shown in Figure 8. E1 represents the surge voltage within an unprotected Ethernet loop. 459 μH is the inductance of the Ethernet loop area with the chassis, 500 pF represents the net series capacitance to ground on both sides of the PSE and PD of the Ethernet connection, and the 10 Ω resistor is the series resistance of the circuit. In simulation, the value of R2 does not actually change the peak value of the current, but causes the envelope of the waveform to decay at a faster rate. This more favorable L/R time constant will allow the surge energy to dissipate as heat more quickly through the distributed resistors.


Figure 8. SPICE simulation model illustrating the ability to reduce inrush current by utilizing a second shorted line turn tightly coupled to the Ethernet loop.


Figure 9. Inrush current from the example simulation in Figure 8.

The generated surge current I(L2) is shown in Figure 9. The graph shows that even if a lightning strike occurs 1 mile away, an unprotected loop will experience a peak-to-peak surge current of 1.6 A. Imagine how much loop current would be generated if the lightning strike was much closer. Even this current is enough to cause damage.

Now, let us consider the inrush current in the protected Ethernet loop (here the inner Ethernet loop) shown in the right half of the schematic. This inrush current can be further reduced if the shield loop impedance is lowered (increasing C3 and C4) while maintaining good magnetic coupling to the Ethernet loop.

isolation

Another way to eliminate inrush current is to isolate one or both ends of the cable. Ideally, to isolate an application in this way, one needs to have an open circuit at all frequencies. This is typically accomplished by an isolation transformer; for Ethernet, this includes data transformers and power transformers (POE applications). Transformers are good at blocking DC; but their primary-to-secondary capacitance shorts out at higher frequencies, supporting high-frequency surge currents. If very low capacitance transformers were available we wouldn't have the surge problem in the first place, so that's not the answer. However, reducing the isolation capacitance does reduce the current caused by lightning strikes. However, the solution proposed in this article provides a better isolation system at higher frequencies, despite the larger capacitance across the isolation barrier. If the capacitor doesn't see any dv/dt, then it doesn't matter.

What is the problem?

The problem is that we can never build ideal shielding around a circuit, or eliminate all magnetic fields with shorted turns, or build a transformer without capacitance. What else can be done in this situation? To enhance these solutions, we may also need to add protection components designed to divert any remaining surge current. The current in the shorted turns can be high, but there's little need to worry since we're only using copper and capacitors to build it. One final improvement we can make is to add ferrite around the entire Ethernet link, as shown in Figure 10.


Figure 10. Common-mode choke CH1 provides low impedance to differential-mode currents and greater impedance to common-mode currents.

This ferrite still performs well without the addition of shorting turns. It provides an open circuit for high frequency currents to supplement the open circuit of the isolation transformer at DC and lower frequencies. If we use ferrites with shorted turns we can get some pretty amazing results. In this case, the ferrite provides an open circuit for current flow around the ground loop, further reducing the C3/C4 ratio.

Conclusion

Any application that requires long cable runs is at risk of damage from lightning strikes. The cause of this damage can be a voltage drop in the ground impedance caused by the high current of a lightning strike (ground loop), as well as a voltage due to Faraday's law (magnetic coupling). In some applications, using protective components to direct this destructive current may not be effective. In this case, adding low impedance shorting turns directly along the Ethernet cable and circuit (with good coupling) can significantly reduce the inrush current. This method uses only copper and capacitors, so we don't have to worry about the high current that shorted turns can create. Adding a common mode choke to your Ethernet cable can also safely reduce inrush current.

References

1 Alan Rich. "Shielding and Protection, How to Eliminate Interfering Noise - Methods and Principles: A Rational Approach." Analog Devices, Inc., 1983.

Karl-Heinz Niemann. "Ethernet-APL Engineering Guide", Version 1.14 19. September 2022.

Richard P. Feynman, Robert B. Leighton and Matthew Sands. Feynman Lectures on Physics, Volume 2: New Millennium Edition: Mainly concerned with electromagnetism and matter. Basic Books, 2011.

Ralph Morrison. Grounding and Shielding Techniques, 4th Edition. John Wiley & Sons Publications, 1998.

About Analog Devices

Analog Devices, Inc. (NASDAQ: ADI) is a leading global semiconductor company dedicated to bridging the physical and digital worlds to enable breakthrough innovations at the intelligent edge. ADI provides solutions that combine analog, digital and software technologies to promote the continued development of digital factories, automobiles and digital health, address the challenges of climate change, and establish reliable interconnections between people and everything in the world. ADI's fiscal year 2023 revenue exceeds US$12 billion and has approximately 26,000 employees worldwide. Working with 125,000 customers around the world, ADI empowers innovators to exceed what is possible.For more information, please visit
www.analog.com/cn

About the author

James Niemann joined Analog Devices in March 2020 and currently works as a field applications engineer in Cleveland, Ohio. James has 35 years of extensive work experience in test and measurement equipment design and currently works as an ADI field application engineer. He holds 14 patents.


#FAQ #ideal #protect #surge #events #Ethernet #industrial #applications